Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method comprising: receiving a depth frame from a depth sensor oriented towards an open end of a shipping container; performing, using a processor, a first occlusion detection analysis on the depth frame to identify a first type of occlusion occurring at a first depth range; performing, using the processor, a second occlusion detection analysis on the depth frame to identify a second type of occlusion occurring at a second depth range different than the first depth range; when one or more occlusions are detected in the depth frame, correcting, using the processor, the one or more occlusions in the depth frame using one or more temporally proximate depth frames; and outputting the corrected depth frame for fullness estimation, wherein the first type of occlusion includes a moving occlusion, and wherein the first occlusion detection analysis comprises identifying that the depth value associated with one of the grid elements decreases with respect to previous frames and then increases in succeeding frames in less than a threshold amount of time across multiple depth frames.
This invention relates to a method for detecting and correcting occlusions in depth frames captured by a depth sensor oriented toward the open end of a shipping container. The method addresses the problem of occlusions, such as moving objects or static obstructions, which can interfere with accurate fullness estimation of the container. The system receives a depth frame from the depth sensor and performs two distinct occlusion detection analyses. The first analysis identifies moving occlusions by detecting depth value fluctuations in grid elements—specifically, when a depth value decreases and then increases within a short timeframe across multiple frames. The second analysis targets a different type of occlusion occurring at a distinct depth range. If occlusions are detected, the method corrects them using temporally proximate depth frames, ensuring accurate depth data. The corrected depth frame is then output for fullness estimation, enabling precise monitoring of container contents. The approach improves reliability in automated inspection systems by mitigating occlusion-related errors.
2. The method of claim 1 , wherein the second type of occlusion includes a missing-data occlusion.
A method for handling occlusions in data processing involves detecting and addressing different types of occlusions that disrupt data integrity. The method focuses on identifying and mitigating missing-data occlusions, where portions of the data are entirely absent or corrupted beyond recovery. This type of occlusion is distinct from other occlusions that may involve partial or temporary disruptions. The method employs techniques to detect missing-data occlusions by analyzing data patterns, statistical anomalies, or metadata inconsistencies. Once identified, the method applies corrective measures such as interpolation, extrapolation, or substitution using alternative data sources to restore the missing information. The approach ensures that the processed data remains accurate and usable despite the presence of missing segments. The method is particularly useful in applications where continuous or complete data is essential, such as sensor networks, medical imaging, or financial data analysis. By distinguishing and addressing missing-data occlusions separately from other types, the method improves the reliability and robustness of data-driven systems.
3. The method of claim 1 , wherein the second occlusion detection analysis comprises: generating a binarization map delineating between (i) grid elements for which the respective depth value is valid and (ii) grid elements for which the respective depth value is not valid; and identifying an instance of the second type of occlusion the occlusion as a cluster of grid elements in the binarization map for which the respective depth value is not valid.
This invention relates to occlusion detection in depth sensing systems, particularly for identifying occlusions in depth maps generated by devices such as LiDAR or structured light sensors. The problem addressed is the accurate detection of occlusions, where certain regions in a scene are obscured, leading to missing or invalid depth values in the depth map. These occlusions can be caused by objects blocking the sensor's line of sight or other environmental factors, resulting in incomplete or erroneous depth data. The method involves analyzing a depth map divided into grid elements, each associated with a depth value. A binarization map is generated to distinguish between valid and invalid depth values. Grid elements with valid depth values are marked as one state, while those with invalid or missing depth values are marked as another. The binarization map is then processed to identify clusters of invalid depth values, which correspond to occluded regions. These clusters are classified as a specific type of occlusion, enabling the system to recognize and handle occluded areas in subsequent processing steps. The approach improves the reliability of depth sensing by explicitly identifying and flagging occluded regions, which can be used for error correction, object tracking, or other applications requiring accurate depth information.
4. The method of claim 3 , wherein identifying the instance of the second type of occlusion comprises confirming that the identified cluster of grid elements exceeds a predetermined occlusion-size threshold.
This invention relates to a method for detecting and classifying occlusions in a grid-based system, such as those used in computer vision or sensor data analysis. The problem addressed is the accurate identification and differentiation of occlusions, which are areas where objects or obstructions block the view or signal path, potentially causing errors in data interpretation. The method involves analyzing a grid composed of multiple elements to detect clusters of elements that indicate an occlusion. The invention specifically focuses on distinguishing between different types of occlusions, such as those caused by objects of varying sizes or characteristics. To confirm the presence of a second type of occlusion, the method checks whether the identified cluster of grid elements exceeds a predetermined size threshold. This threshold ensures that only significant occlusions are flagged, filtering out minor or insignificant obstructions that may not affect system performance. The method may also include preprocessing steps to prepare the grid data, such as filtering noise or enhancing relevant features, to improve occlusion detection accuracy. By setting a size-based threshold, the system can reliably classify occlusions, enabling more robust decision-making in applications like autonomous navigation, surveillance, or environmental monitoring. The approach ensures that only meaningful occlusions are identified, reducing false positives and improving overall system reliability.
5. The method of claim 4 , wherein identifying the instance of the second type of occlusion further comprises performing edge detection on the cluster of grid elements.
This invention relates to a method for detecting occlusions in a grid-based system, particularly for identifying and classifying different types of occlusions in a structured grid environment. The method addresses the challenge of accurately distinguishing between occlusions caused by different factors, such as physical obstructions or sensor failures, in applications like autonomous navigation, robotics, or environmental monitoring. The method involves analyzing a grid composed of multiple grid elements to detect occlusions. Initially, a cluster of grid elements is identified where an occlusion is suspected. To refine the detection, edge detection techniques are applied to the cluster. Edge detection helps in identifying boundaries or transitions within the cluster, which are indicative of the type of occlusion present. For example, sharp edges may suggest a physical obstruction, while diffuse edges may indicate sensor-related issues. The method builds upon a prior step of classifying the occlusion into a first type, such as a physical obstruction, and then further refining the classification by performing edge detection on the cluster. This additional step enhances the accuracy of occlusion identification, allowing for more precise decision-making in systems that rely on grid-based data. By incorporating edge detection, the method improves the reliability of occlusion detection in dynamic environments, ensuring that systems can adapt to varying conditions and maintain operational integrity. This approach is particularly useful in applications where real-time detection and classification of occlusions are critical for safety and efficiency.
6. The method of claim 5 , wherein identifying the instance of the second type of occlusion further comprises performing contour identification on the cluster of grid elements.
This invention relates to a method for detecting and analyzing occlusions in a visual scene, particularly for identifying different types of occlusions in a grid-based representation of the scene. The method addresses the challenge of accurately distinguishing between occlusions caused by different objects or structures, which is critical in applications such as autonomous navigation, object recognition, and scene reconstruction. The method involves processing a grid-based representation of the scene, where the grid is divided into elements that may be partially or fully occluded. A cluster of grid elements is identified, representing a region where occlusion is likely present. To further refine the detection, the method performs contour identification on the cluster of grid elements. This step involves analyzing the boundaries and shapes of the occluded region to determine whether the occlusion is of a second type, distinct from a first type previously identified. The contour identification helps differentiate between occlusions caused by different objects or environmental factors, improving the accuracy of the analysis. By performing contour identification, the method enhances the ability to classify occlusions, which is useful in applications requiring precise environmental understanding, such as robotics, augmented reality, and computer vision systems. The method ensures that occlusions are not only detected but also accurately categorized, enabling better decision-making in automated systems.
7. The method of claim 1 , wherein the moving occlusion is associated with a single grid element in the plurality of grid elements.
A method for analyzing moving occlusions in a grid-based system involves tracking occlusions that obstruct visibility or movement within a defined grid structure. The grid is divided into multiple grid elements, each representing a discrete spatial region. The method detects and associates a moving occlusion with a single grid element, ensuring precise localization of the obstruction within the grid. This association allows for accurate tracking of the occlusion's position and movement as it transitions between grid elements. The method may also include determining the occlusion's impact on visibility or movement within the grid, such as blocking lines of sight or pathways. By restricting the occlusion to a single grid element at any given time, the method simplifies the analysis of dynamic obstructions in grid-based environments, such as in robotics, autonomous navigation, or computer vision applications. The technique improves spatial awareness and collision avoidance by maintaining a clear and discrete representation of moving obstructions within the grid framework.
8. The method of claim 1 , wherein the first occlusion detection analysis further comprises identifying a threshold depth change in one of the grid elements between the depth frame and at least one temporally proximate depth frame.
This invention relates to depth-based occlusion detection in imaging systems, particularly for identifying occlusions in dynamic scenes. The method involves analyzing depth frames to detect occlusions by comparing depth values across multiple frames. A key aspect is identifying a threshold depth change in specific grid elements of the depth frame when compared to at least one temporally proximate depth frame. This helps distinguish between true occlusions and noise or minor depth variations. The method may also involve segmenting the depth frame into a grid of elements, where each element represents a portion of the scene. By tracking depth changes within these grid elements over time, the system can accurately detect when an object or part of the scene is occluded. The threshold depth change ensures that only significant occlusions are flagged, reducing false positives. This approach is useful in applications like robotics, augmented reality, and surveillance, where reliable occlusion detection is critical for accurate scene understanding and interaction. The method improves upon existing techniques by providing a more robust and precise way to identify occlusions in dynamic environments.
9. The method of claim 1 , wherein the first type of occlusion further includes a discontinuous occlusion.
A method for managing occlusions in a fluid delivery system, particularly in medical or industrial applications where precise fluid control is critical. The method addresses the challenge of detecting and mitigating occlusions that disrupt fluid flow, which can lead to system failures or inaccurate dosing. The invention focuses on identifying and handling different types of occlusions, including continuous and discontinuous occlusions, to ensure reliable operation. The method involves monitoring fluid flow within a system to detect occlusions. A discontinuous occlusion is characterized by intermittent or partial blockages that allow some fluid to pass but disrupt normal flow dynamics. This type of occlusion can be caused by debris, air bubbles, or partial obstructions in the fluid path. The method distinguishes discontinuous occlusions from continuous occlusions, which completely block fluid flow, and applies specific mitigation strategies for each. Upon detecting a discontinuous occlusion, the method may adjust system parameters such as pressure, flow rate, or valve settings to restore proper fluid delivery. The system may also log the occurrence for maintenance or diagnostic purposes. The method ensures that fluid delivery remains accurate and consistent, even in the presence of intermittent obstructions, improving system reliability and safety.
10. The method of claim 9 , wherein the first occlusion detection analysis identifying the first type of occlusion comprises: identifying a cluster of grid elements having a collective depth value; calculating a difference between the collective depth value and a depth value of a loaded-portion boundary of the shipping container; comparing the difference to a threshold; and identifying the discontinuous occlusion when the difference is greater than the threshold.
11. The method of claim 10 , wherein identifying the discontinuous occlusion further comprises confirming that the identified cluster of grid elements exceeds a predetermined occlusion-size threshold.
This invention relates to a method for detecting discontinuous occlusions in a grid-based system, such as those used in computer vision, medical imaging, or autonomous navigation. The problem addressed is the accurate identification of occlusions that are not continuous, meaning they may appear as fragmented or irregularly shaped regions within the grid. Such occlusions can disrupt data processing, object recognition, or navigation decisions if not properly detected. The method involves analyzing a grid composed of multiple elements to identify clusters of elements that represent an occlusion. Once a cluster is detected, the system confirms whether the cluster meets a predetermined size threshold. This threshold ensures that only significant occlusions are flagged, filtering out minor or insignificant disruptions that may not impact system performance. The size threshold is a configurable parameter, allowing the method to adapt to different applications where the definition of a "significant" occlusion may vary. By enforcing this threshold, the method improves the reliability of occlusion detection, reducing false positives and ensuring that only meaningful occlusions are processed further. This approach is particularly useful in dynamic environments where occlusions may be transient or partially obscured.
12. The method of claim 11 , wherein identifying the discontinuous occlusion further comprises performing edge detection on the cluster of grid elements.
13. The method of claim 12 , wherein identifying the discontinuous occlusion further comprises performing contour identification on the cluster of grid elements.
14. The method of claim 1 , wherein the grid elements are pixels.
A system and method for processing image data involves analyzing a grid of elements to detect and correct distortions or errors. The grid elements are pixels, each representing a discrete unit of the image. The method includes capturing or receiving an image composed of these pixels, where each pixel contains color or intensity values. The system then applies a transformation or correction algorithm to the pixel grid to improve image quality, such as reducing noise, enhancing resolution, or correcting geometric distortions. The transformation may involve interpolation, filtering, or other computational techniques to adjust pixel values based on neighboring pixels or predefined criteria. The corrected pixel grid is then output as a refined image. This approach is useful in applications like medical imaging, surveillance, or digital photography, where accurate pixel-level processing is critical for reliable analysis or display. The method ensures that distortions at the pixel level are minimized, improving overall image fidelity.
15. The method of claim 1 , wherein the grid elements are groups of pixels.
A method for processing image data involves organizing pixels into grid elements, where each grid element consists of a group of pixels. The method includes capturing an image using an imaging device, such as a camera, and dividing the captured image into a grid structure. Each grid element within this structure represents a localized region of the image, containing multiple pixels. The method further processes these grid elements to extract features, analyze patterns, or perform other image-related tasks. The grouping of pixels into grid elements allows for efficient data handling, reducing computational complexity while maintaining spatial relationships within the image. This approach is particularly useful in applications like object detection, image segmentation, or feature extraction, where managing pixel data in structured groups improves processing speed and accuracy. The method may also include additional steps such as filtering, normalization, or transformation of the grid elements to enhance image analysis. By organizing pixels into grid elements, the method optimizes image processing workflows, making it suitable for real-time applications or systems with limited computational resources.
16. The method of claim 1 , wherein: the one or more identified occlusions corresponds to an occlusion set of the grid elements in the depth frame; and correcting the one or more occlusions in the depth frame using one or more temporally proximate depth frames comprises overwriting the occlusion set in the depth frame with data from corresponding non-occluded grid elements from one or more of the temporally proximate depth frames.
17. The method of claim 1 , further comprising analyzing a buffer of depth frames, wherein the buffer includes the received depth frame.
Unknown
July 14, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.